Authors:
Eduardo Augusto Klosowski
and
Adriano Fiorese
Affiliation:
Graduate Program in Applied Computing (PPGCA), Dept. of Computer Science (DCC), Santa Catarina State University (UDESC), Joinville and Brazil
Keyword(s):
Evaluation Framework, SDN Controller, Relative Efficiency, DEA.
Related
Ontology
Subjects/Areas/Topics:
Agents
;
Artificial Intelligence
;
Artificial Intelligence and Decision Support Systems
;
Bioinformatics
;
Biomedical Engineering
;
Enterprise Information Systems
;
Information Systems Analysis and Specification
;
Internet Systems Performance
;
Knowledge Discovery and Information Retrieval
;
Knowledge-Based Systems
;
Methodologies and Technologies
;
Operational Research
;
Problem Solving
;
Simulation
;
Software Agents and Internet Computing
;
Software Metrics and Measurement
;
Strategic Decision Support Systems
;
Symbolic Systems
;
Tools, Techniques and Methodologies for System Development
;
User Profiling and Recommender Systems
Abstract:
A Software-Defined Network (SDN) requires a controller that is responsible for defining how the network will behave, since it has the responsibility to install flow rules for forwarding the data streams through the network devices. Thus, it is necessary that the controller presents performance good enough to attend the network needs. However, with the diversity of existing controllers, some offering more facilities to the developer, while others offer higher performance, a doubt arises regarding which controller manages to attend the network demand, or how much performance can be bargained to get more facilities. To answer these questions, this work presents FREEController. It is an SDN controllers evaluation framework based on relative efficiency obtained by means the Data Envelopment Analysis (DEA) multicriteria decision-making method. This proposed framework takes into account several stages including the controllers’ performance evaluation, creating a performance database, and ho
w to use these database to identify which controllers attend the network demand using the DEA method. Results comprising the proposed framework evaluation indicate the viability of the relative efficiency approach and its relation with the used controllers’ resources.
(More)